| 6 Aug 2025 |
 j-k | Is discourse.nixos.org down? I'm getting 502 errors from nginx and isitdownrightnow.com seems to get the same result | 19:21:03 |
 sorrel | mhm, also seems very slow to me when it responds at all | 19:24:20 |
 [0x4A6F] | Still up here. | 19:27:49 |
| [0x4A6F] | * Still up here. With banner: Due to extreme load, this is temporarily being shown to everyone as a logged out user would see it. | 19:28:29 |
 Christian Theune | yeah, we got hit by the same ceph bug as yesterday. i've found a trigger that we can stop from triggering for now so this shouldn't become a regular thing ... -_- | 19:36:50 |
| Christian Theune | things should be back to normal | 19:37:07 |
| 7 Aug 2025 |
 hexa | biweekly infra call in 25m https://meet.cccda.de/tzi-rabl-tro | 15:37:04 |
| 8 Aug 2025 |
|  sinan changed their profile picture. | 06:11:32 |
|  0x4A6F joined the room. | 06:59:01 |
|  Picnoir changed their display name from Picnoir to Picnoir DECT 7426. | 12:32:31 |
| 9 Aug 2025 |
|  dmitrii_s joined the room. | 19:06:57 |
| 10 Aug 2025 |
| hexa | ok, so as discussed in #infra-alerts:nixos.org we want to replace haumea.nixos.org, which has been showing (correctable) ecc errors for a while now | 21:58:41 |
| hexa | the plan is for an ax52 with the ecc memory upgrade | 21:59:25 |
| 11 Aug 2025 |
 @emma:rory.gay | woah thats pretty expensive for the specs | 11:46:30 |
| hexa | compared to what | 11:46:59 |
| @emma:rory.gay | the server i have at ovh for ~20 EUR more (a KS-6: https://eco.ovhcloud.com/en/kimsufi/ks-6/) | 11:48:17 |
| @emma:rory.gay | (though mine is with the +~35 EUR 256GiB ram upgrade...) | 11:49:13 |
| hexa | fair, but we are locked into the location due to latency requirements | 11:49:18 |
| hexa | this machine hosts the database for hydra.nixos.org | 11:49:38 |
| @emma:rory.gay | ah, im guessing an OVH DC wouldnt be close enough | 11:49:57 |
| @emma:rory.gay | * ah, im guessing an OVH DC wouldnt be close enough for that, network overhead wise | 11:51:22 |
 Vladimír Čunát | OVH says to me that this server isn't available (except with drives that are too small for us). | 11:58:11 |
 emily | the Kimsufi stuff has a bunch of somewhat awkward limitations, I think? (dunno how much the 500 Mbit/s pipe there would matter for a database server but I imagine it would be quite painful for many Hydra cluster machines) | 12:03:29 |
| Vladimír Čunát | Ah. We sometimes do use higher traffic than that on this particular server. | 12:09:28 |
| @emma:rory.gay | tbh i'd say that's fine if its peak load rather than sustained | 12:09:49 |
| Vladimír Čunát | * Ah. We sometimes do use higher traffic than that on this particular server (towards our other server). | 12:10:09 |
| @emma:rory.gay | also, damn, they changed the network offering since i got mine lol (im limited to 300 symmetrical) | 12:10:26 |
| @emma:rory.gay | im just dealing with the stock 2x500gb (without raid as i needed the extra space, but didnt want the perfomrnace impact of using HDDs) | 12:11:23 |
| hexa | our database is currently at 478G + snapshots | 12:13:43 |
| hexa | currently on 2x1TB and that seems sufficient for now | 12:14:15 |
